/* Currently these are only used when PL_regkind[OP(rn)] == EXACT so
we don't need this definition. */
#define IS_TEXT(rn) ( OP(rn)==EXACT || OP(rn)==REF || OP(rn)==NREF )
-#define IS_TEXTF(rn) ( OP(rn)==EXACTFU || OP(rn)==EXACTFU_SS || OP(rn)==EXACTFA || OP(rn)==EXACTF || OP(rn)==REFF || OP(rn)==NREFF )
+#define IS_TEXTF(rn) ( OP(rn)==EXACTFU || OP(rn)==EXACTFU_SS || OP(rn)==EXACTFU_TRICKYFOLD || OP(rn)==EXACTFA || OP(rn)==EXACTF || OP(rn)==REFF || OP(rn)==NREFF )
#define IS_TEXTFL(rn) ( OP(rn)==EXACTFL || OP(rn)==REFFL || OP(rn)==NREFFL )
#else
/* ... so we use this as its faster. */
#define IS_TEXT(rn) ( OP(rn)==EXACT )
-#define IS_TEXTFU(rn) ( OP(rn)==EXACTFU || OP(rn)==EXACTFU_SS || OP(rn) == EXACTFA)
+#define IS_TEXTFU(rn) ( OP(rn)==EXACTFU || OP(rn)==EXACTFU_SS || OP(rn)==EXACTFU_TRICKYFOLD || OP(rn) == EXACTFA)
#define IS_TEXTF(rn) ( OP(rn)==EXACTF )
#define IS_TEXTFL(rn) ( OP(rn)==EXACTFL )
GET_RE_DEBUG_FLAGS_DECL;
PERL_ARGS_ASSERT_RE_INTUIT_START;
+ PERL_UNUSED_ARG(flags);
+ PERL_UNUSED_ARG(data);
RX_MATCH_UTF8_set(rx,utf8_target);
{
I32 srch_start_shift = start_shift;
I32 srch_end_shift = end_shift;
+ U8* start_point;
+ U8* end_point;
if (srch_start_shift < 0 && strbeg - s > srch_start_shift) {
srch_end_shift -= ((strbeg - s) - srch_start_shift);
srch_start_shift = strbeg - s;
(IV)prog->check_end_shift);
});
- if ((flags & REXEC_SCREAM) && SvSCREAM(sv)) {
- I32 p = -1; /* Internal iterator of scream. */
- I32 * const pp = data ? data->scream_pos : &p;
- const MAGIC *mg;
- bool found = FALSE;
-
- assert(SvMAGICAL(sv));
- mg = mg_find(sv, PERL_MAGIC_study);
- assert(mg);
-
- if (mg->mg_private == 1) {
- found = ((U8 *)mg->mg_ptr)[BmRARE(check)] != (U8)~0;
- } else if (mg->mg_private == 2) {
- found = ((U16 *)mg->mg_ptr)[BmRARE(check)] != (U16)~0;
- } else {
- assert (mg->mg_private == 4);
- found = ((U32 *)mg->mg_ptr)[BmRARE(check)] != (U32)~0;
- }
-
- if (found
- || ( BmRARE(check) == '\n'
- && (BmPREVIOUS(check) == SvCUR(check) - 1)
- && SvTAIL(check) ))
- s = screaminstr(sv, check,
- srch_start_shift + (s - strbeg), srch_end_shift, pp, 0);
- else
- goto fail_finish;
- /* we may be pointing at the wrong string */
- if (s && RXp_MATCH_COPIED(prog))
- s = strbeg + (s - SvPVX_const(sv));
- if (data)
- *data->scream_olds = s;
- }
- else {
- U8* start_point;
- U8* end_point;
if (prog->extflags & RXf_CANY_SEEN) {
start_point= (U8*)(s + srch_start_shift);
end_point= (U8*)(strend - srch_end_shift);
s = fbm_instr( start_point, end_point,
check, multiline ? FBMrf_MULTILINE : 0);
}
- }
/* Update the count-of-usability, remove useless subpatterns,
unshift s. */
#define DECL_TRIE_TYPE(scan) \
const enum { trie_plain, trie_utf8, trie_utf8_fold, trie_latin_utf8_fold } \
- trie_type = (scan->flags != EXACT) \
- ? (utf8_target ? trie_utf8_fold : (UTF_PATTERN ? trie_latin_utf8_fold : trie_plain)) \
- : (utf8_target ? trie_utf8 : trie_plain)
-
-#define REXEC_TRIE_READ_CHAR(trie_type, trie, widecharmap, uc, uscan, len, \
-uvc, charid, foldlen, foldbuf, uniflags) STMT_START { \
- switch (trie_type) { \
- case trie_utf8_fold: \
- if ( foldlen>0 ) { \
- uvc = utf8n_to_uvuni( uscan, UTF8_MAXLEN, &len, uniflags ); \
- foldlen -= len; \
- uscan += len; \
- len=0; \
- } else { \
- uvc = to_utf8_fold( (U8 *) uc, foldbuf, &foldlen ); \
- len = UTF8SKIP(uc); \
- foldlen -= UNISKIP( uvc ); \
- uscan = foldbuf + UNISKIP( uvc ); \
- } \
- break; \
- case trie_latin_utf8_fold: \
- if ( foldlen>0 ) { \
- uvc = utf8n_to_uvuni( uscan, UTF8_MAXLEN, &len, uniflags ); \
- foldlen -= len; \
- uscan += len; \
- len=0; \
- } else { \
- len = 1; \
- uvc = to_uni_fold( *(U8*)uc, foldbuf, &foldlen ); \
- foldlen -= UNISKIP( uvc ); \
- uscan = foldbuf + UNISKIP( uvc ); \
- } \
- break; \
- case trie_utf8: \
- uvc = utf8n_to_uvuni( (U8*)uc, UTF8_MAXLEN, &len, uniflags ); \
- break; \
- case trie_plain: \
- uvc = (UV)*uc; \
- len = 1; \
- } \
- if (uvc < 256) { \
- charid = trie->charmap[ uvc ]; \
- } \
- else { \
- charid = 0; \
- if (widecharmap) { \
- SV** const svpp = hv_fetch(widecharmap, \
- (char*)&uvc, sizeof(UV), 0); \
- if (svpp) \
- charid = (U16)SvIV(*svpp); \
- } \
- } \
+ trie_type = ((scan->flags == EXACT) \
+ ? (utf8_target ? trie_utf8 : trie_plain) \
+ : (utf8_target ? trie_utf8_fold : trie_latin_utf8_fold))
+
+#define REXEC_TRIE_READ_CHAR(trie_type, trie, widecharmap, uc, uscan, len, \
+uvc, charid, foldlen, foldbuf, uniflags) STMT_START { \
+ STRLEN skiplen; \
+ switch (trie_type) { \
+ case trie_utf8_fold: \
+ if ( foldlen>0 ) { \
+ uvc = utf8n_to_uvuni( (const U8*) uscan, UTF8_MAXLEN, &len, uniflags ); \
+ foldlen -= len; \
+ uscan += len; \
+ len=0; \
+ } else { \
+ uvc = to_utf8_fold( (const U8*) uc, foldbuf, &foldlen ); \
+ len = UTF8SKIP(uc); \
+ skiplen = UNISKIP( uvc ); \
+ foldlen -= skiplen; \
+ uscan = foldbuf + skiplen; \
+ } \
+ break; \
+ case trie_latin_utf8_fold: \
+ if ( foldlen>0 ) { \
+ uvc = utf8n_to_uvuni( (const U8*) uscan, UTF8_MAXLEN, &len, uniflags ); \
+ foldlen -= len; \
+ uscan += len; \
+ len=0; \
+ } else { \
+ len = 1; \
+ uvc = _to_fold_latin1( (U8) *uc, foldbuf, &foldlen, 1); \
+ skiplen = UNISKIP( uvc ); \
+ foldlen -= skiplen; \
+ uscan = foldbuf + skiplen; \
+ } \
+ break; \
+ case trie_utf8: \
+ uvc = utf8n_to_uvuni( (const U8*) uc, UTF8_MAXLEN, &len, uniflags ); \
+ break; \
+ case trie_plain: \
+ uvc = (UV)*uc; \
+ len = 1; \
+ } \
+ if (uvc < 256) { \
+ charid = trie->charmap[ uvc ]; \
+ } \
+ else { \
+ charid = 0; \
+ if (widecharmap) { \
+ SV** const svpp = hv_fetch(widecharmap, \
+ (char*)&uvc, sizeof(UV), 0); \
+ if (svpp) \
+ charid = (U16)SvIV(*svpp); \
+ } \
+ } \
} STMT_END
#define REXEC_FBC_EXACTISH_SCAN(CoNd) \
}
goto do_exactf_utf8;
+ case EXACTFU_TRICKYFOLD:
case EXACTFU:
if (UTF_PATTERN || utf8_target) {
utf8_fold_flags = (UTF_PATTERN) ? FOLDEQ_S2_ALREADY_FOLDED : 0;
/*XXX: The s-- is almost definitely wrong here under unicode - demeprhq*/
s--;
}
- /* We can use a more efficient search as newlines are the same in unicode as they are in latin */
- while (s < end) {
+ /* We can use a more efficient search as newlines are the same in unicode as they are in latin */
+ while (s <= end) { /* note it could be possible to match at the end of the string */
if (*s++ == '\n') { /* don't need PL_utf8skip here */
if (regtry(®info, &s))
goto got_it;
dontbother = end_shift;
strend = HOPc(strend, -dontbother);
while ( (s <= last) &&
- ((flags & REXEC_SCREAM) && SvSCREAM(sv)
- ? (s = screaminstr(sv, must, HOP3c(s, back_min, (back_min<0 ? strbeg : strend)) - strbeg,
- end_shift, &scream_pos, 0))
- : (s = fbm_instr((unsigned char*)HOP3(s, back_min, (back_min<0 ? strbeg : strend)),
+ (s = fbm_instr((unsigned char*)HOP3(s, back_min, (back_min<0 ? strbeg : strend)),
(unsigned char*)strend, must,
- multiline ? FBMrf_MULTILINE : 0))) ) {
- /* we may be pointing at the wrong string */
- if ((flags & REXEC_SCREAM) && RXp_MATCH_COPIED(prog))
- s = strbeg + (s - SvPVX_const(sv));
+ multiline ? FBMrf_MULTILINE : 0)) ) {
DEBUG_EXECUTE_r( did_match = 1 );
if (HOPc(s, -back_max) > last1) {
last1 = HOPc(s, -back_min);
dontbother = 0;
if (prog->float_substr != NULL || prog->float_utf8 != NULL) {
/* Trim the end. */
- char *last;
+ char *last= NULL;
SV* float_real;
+ STRLEN len;
+ const char *little;
if (!(utf8_target ? prog->float_utf8 : prog->float_substr))
utf8_target ? to_utf8_substr(prog) : to_byte_substr(prog);
float_real = utf8_target ? prog->float_utf8 : prog->float_substr;
- if ((flags & REXEC_SCREAM) && SvSCREAM(sv)) {
- last = screaminstr(sv, float_real, s - strbeg,
- end_shift, &scream_pos, 1); /* last one */
- if (!last)
- last = scream_olds; /* Only one occurrence. */
- /* we may be pointing at the wrong string */
- else if (RXp_MATCH_COPIED(prog))
- s = strbeg + (s - SvPVX_const(sv));
- }
- else {
- STRLEN len;
- const char * const little = SvPV_const(float_real, len);
-
- if (SvTAIL(float_real)) {
- if (memEQ(strend - len + 1, little, len - 1))
- last = strend - len + 1;
- else if (!multiline)
- last = memEQ(strend - len, little, len)
- ? strend - len : NULL;
- else
+ little = SvPV_const(float_real, len);
+ if (SvTAIL(float_real)) {
+ /* This means that float_real contains an artificial \n on the end
+ * due to the presence of something like this: /foo$/
+ * where we can match both "foo" and "foo\n" at the end of the string.
+ * So we have to compare the end of the string first against the float_real
+ * without the \n and then against the full float_real with the string.
+ * We have to watch out for cases where the string might be smaller
+ * than the float_real or the float_real without the \n.
+ */
+ char *checkpos= strend - len;
+ DEBUG_OPTIMISE_r(
+ PerlIO_printf(Perl_debug_log,
+ "%sChecking for float_real.%s\n",
+ PL_colors[4], PL_colors[5]));
+ if (checkpos + 1 < strbeg) {
+ /* can't match, even if we remove the trailing \n string is too short to match */
+ DEBUG_EXECUTE_r(
+ PerlIO_printf(Perl_debug_log,
+ "%sString shorter than required trailing substring, cannot match.%s\n",
+ PL_colors[4], PL_colors[5]));
+ goto phooey;
+ } else if (memEQ(checkpos + 1, little, len - 1)) {
+ /* can match, the end of the string matches without the "\n" */
+ last = checkpos + 1;
+ } else if (checkpos < strbeg) {
+ /* cant match, string is too short when the "\n" is included */
+ DEBUG_EXECUTE_r(
+ PerlIO_printf(Perl_debug_log,
+ "%sString does not contain required trailing substring, cannot match.%s\n",
+ PL_colors[4], PL_colors[5]));
+ goto phooey;
+ } else if (!multiline) {
+ /* non multiline match, so compare with the "\n" at the end of the string */
+ if (memEQ(checkpos, little, len)) {
+ last= checkpos;
+ } else {
+ DEBUG_EXECUTE_r(
+ PerlIO_printf(Perl_debug_log,
+ "%sString does not contain required trailing substring, cannot match.%s\n",
+ PL_colors[4], PL_colors[5]));
+ goto phooey;
+ }
+ } else {
+ /* multiline match, so we have to search for a place where the full string is located */
goto find_last;
- } else {
+ }
+ } else {
find_last:
if (len)
last = rninstr(s, strend, little, little + len);
else
last = strend; /* matching "$" */
- }
}
- if (last == NULL) {
+ if (!last) {
+ /* at one point this block contained a comment which was probably
+ * incorrect, which said that this was a "should not happen" case.
+ * Even if it was true when it was written I am pretty sure it is
+ * not anymore, so I have removed the comment and replaced it with
+ * this one. Yves */
DEBUG_EXECUTE_r(
PerlIO_printf(Perl_debug_log,
- "%sCan't trim the tail, match fails (should not happen)%s\n",
- PL_colors[4], PL_colors[5]));
- goto phooey; /* Should not happen! */
+ "String does not contain required substring, cannot match.\n"
+ ));
+ goto phooey;
}
dontbother = strend - last + prog->float_min_offset;
}
/* In this case the charclass data is available inline so
we can fail fast without a lot of extra overhead.
*/
- if (scan->flags == EXACT || !utf8_target) {
- if(!ANYOF_BITMAP_TEST(scan, *locinput)) {
- DEBUG_EXECUTE_r(
- PerlIO_printf(Perl_debug_log,
- "%*s %sfailed to match trie start class...%s\n",
- REPORT_CODE_OFF+depth*2, "", PL_colors[4], PL_colors[5])
- );
- sayNO_SILENT;
- /* NOTREACHED */
- }
+ if(!ANYOF_BITMAP_TEST(scan, *locinput)) {
+ DEBUG_EXECUTE_r(
+ PerlIO_printf(Perl_debug_log,
+ "%*s %sfailed to match trie start class...%s\n",
+ REPORT_CODE_OFF+depth*2, "", PL_colors[4], PL_colors[5])
+ );
+ sayNO_SILENT;
+ /* NOTREACHED */
}
/* FALL THROUGH */
case TRIE:
HV * widecharmap = MUTABLE_HV(rexi->data->data[ ARG( scan ) + 1 ]);
U32 state = trie->startstate;
- if (trie->bitmap && trie_type != trie_utf8_fold &&
- !TRIE_BITMAP_TEST(trie,*locinput)
- ) {
+ if (trie->bitmap && !TRIE_BITMAP_TEST(trie,*locinput) ) {
if (trie->states[ state ].wordnum) {
DEBUG_EXECUTE_r(
PerlIO_printf(Perl_debug_log,
goto do_exactf;
case EXACTFU_SS:
+ case EXACTFU_TRICKYFOLD:
case EXACTFU:
folder = foldEQ_latin1;
fold_array = PL_fold_latin1;
case EXACTF: ST.c2 = PL_fold[ST.c1]; break;
case EXACTFA:
case EXACTFU_SS:
+ case EXACTFU_TRICKYFOLD:
case EXACTFU: ST.c2 = PL_fold_latin1[ST.c1]; break;
case EXACTFL: ST.c2 = PL_fold_locale[ST.c1]; break;
default: ST.c2 = ST.c1;
case EXACTF: ST.c2 = PL_fold[ST.c1]; break;
case EXACTFA:
case EXACTFU_SS:
+ case EXACTFU_TRICKYFOLD:
case EXACTFU: ST.c2 = PL_fold_latin1[ST.c1]; break;
case EXACTFL: ST.c2 = PL_fold_locale[ST.c1]; break;
default: ST.c2 = ST.c1; break;
}
else { /* UTF_PATTERN */
if (IS_TEXTFU(text_node) || IS_TEXTF(text_node)) {
- STRLEN ulen1, ulen2;
- U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
- U8 tmpbuf2[UTF8_MAXBYTES_CASE+1];
+ STRLEN ulen;
+ U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
- to_utf8_lower((U8*)s, tmpbuf1, &ulen1);
- to_utf8_upper((U8*)s, tmpbuf2, &ulen2);
-#ifdef EBCDIC
- ST.c1 = utf8n_to_uvchr(tmpbuf1, UTF8_MAXLEN, 0,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY);
- ST.c2 = utf8n_to_uvchr(tmpbuf2, UTF8_MAXLEN, 0,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY);
-#else
- ST.c1 = utf8n_to_uvuni(tmpbuf1, UTF8_MAXBYTES, 0,
- uniflags);
- ST.c2 = utf8n_to_uvuni(tmpbuf2, UTF8_MAXBYTES, 0,
+ to_utf8_fold((U8*)s, tmpbuf, &ulen);
+ ST.c1 = ST.c2 = utf8n_to_uvchr(tmpbuf, UTF8_MAXLEN, 0,
uniflags);
-#endif
}
else {
ST.c2 = ST.c1 = utf8n_to_uvchr(s, UTF8_MAXBYTES, 0,
sayNO;
/* NOTREACHED */
#undef ST
- case FOLDCHAR:
- n = ARG(scan);
- if ( n == (U32)what_len_TRICKYFOLD(locinput,utf8_target,ln) ) {
- locinput += ln;
- } else if ( LATIN_SMALL_LETTER_SHARP_S == n && !utf8_target && !UTF_PATTERN ) {
- sayNO;
- } else {
- U8 folded[UTF8_MAXBYTES_CASE+1];
- STRLEN foldlen;
- const char * const l = locinput;
- char *e = PL_regeol;
- to_uni_fold(n, folded, &foldlen);
-
- if (! foldEQ_utf8((const char*) folded, 0, foldlen, 1,
- l, &e, 0, utf8_target)) {
- sayNO;
- }
- locinput = e;
- }
- nextchr = UCHARAT(locinput);
- break;
case LNBREAK:
if ((n=is_LNBREAK(locinput,utf8_target))) {
locinput += n;
goto do_exactf;
case EXACTFU_SS:
+ case EXACTFU_TRICKYFOLD:
case EXACTFU:
utf8_flags = (UTF_PATTERN) ? FOLDEQ_S2_ALREADY_FOLDED : 0;
switch (OP(p)) {
case EXACTF: folded = PL_fold[c]; break;
case EXACTFA:
+ case EXACTFU_TRICKYFOLD:
case EXACTFU: folded = PL_fold_latin1[c]; break;
case EXACTFL: folded = PL_fold_locale[c]; break;
default: Perl_croak(aTHX_ "panic: Unexpected op %u", OP(p));
(ANYOF_CLASS_TEST(n, ANYOF_NALNUMC) && !isALNUMC_LC(c)) ||
(ANYOF_CLASS_TEST(n, ANYOF_ALPHA) && isALPHA_LC(c)) ||
(ANYOF_CLASS_TEST(n, ANYOF_NALPHA) && !isALPHA_LC(c)) ||
- (ANYOF_CLASS_TEST(n, ANYOF_ASCII) && isASCII(c)) ||
- (ANYOF_CLASS_TEST(n, ANYOF_NASCII) && !isASCII(c)) ||
+ (ANYOF_CLASS_TEST(n, ANYOF_ASCII) && isASCII_LC(c)) ||
+ (ANYOF_CLASS_TEST(n, ANYOF_NASCII) && !isASCII_LC(c)) ||
(ANYOF_CLASS_TEST(n, ANYOF_CNTRL) && isCNTRL_LC(c)) ||
(ANYOF_CLASS_TEST(n, ANYOF_NCNTRL) && !isCNTRL_LC(c)) ||
(ANYOF_CLASS_TEST(n, ANYOF_GRAPH) && isGRAPH_LC(c)) ||
(ANYOF_CLASS_TEST(n, ANYOF_NXDIGIT) && !isXDIGIT(c)) ||
(ANYOF_CLASS_TEST(n, ANYOF_PSXSPC) && isPSXSPC(c)) ||
(ANYOF_CLASS_TEST(n, ANYOF_NPSXSPC) && !isPSXSPC(c)) ||
- (ANYOF_CLASS_TEST(n, ANYOF_BLANK) && isBLANK(c)) ||
- (ANYOF_CLASS_TEST(n, ANYOF_NBLANK) && !isBLANK(c))
+ (ANYOF_CLASS_TEST(n, ANYOF_BLANK) && isBLANK_LC(c)) ||
+ (ANYOF_CLASS_TEST(n, ANYOF_NBLANK) && !isBLANK_LC(c))
) /* How's that for a conditional? */
) {
match = TRUE;
* Local variables:
* c-indentation-style: bsd
* c-basic-offset: 4
- * indent-tabs-mode: t
+ * indent-tabs-mode: nil
* End:
*
- * ex: set ts=8 sts=4 sw=4 noet:
+ * ex: set ts=8 sts=4 sw=4 et:
*/
https://perl5.git.perl.org / perl5.git / blobdiff